Like ... just about everything else in initial 4e the problem is shown most clearly by the Ranger (those that aren't are shown by the Paladin).
At Will: Twin Strike
Encounter 1: Fox's Retreat
Encounter 3: Disruptive Strike
Both the encounter powers are very easy to trigger. The Ranger can therefore fire six arrows in the first two rounds, and therefore multiplies their static damage bonus by 3 every fight at the time you most need damage. With just the PHB this is already troubling - it does too much damage out of the gate even before it scales far too well with "just" Weapon Focus and +X weapons. Throw in Adventurer's Vault for a greatbow and Bracers of Archery and things get a whole lot worse.
I'm not implying it was 'power creep' in later books that is the problem. I think that the problem is that designers of a game imagine how it will work in their heads. They come up with a mechanic and in isolation its a perfectly good thing, like Disruptive Strike seems like a cool thing to give to the ranger. They think "Oh, yeah, it will up his damage a bit, but he's a striker, that's acceptable" and eventually the whole book gets put together and the other little decisions "Twin Strike will be OK for the ranger, its just a little bit of extra damage" and all the little ways of getting a static bonus that the feat writing guys added in "Oh, weapon focus is just an extra point of damage per attack, that's trivial", etc. Now you bring it all together in totality and its not so good. However the designer's still have the idea of the game fixed in their heads that Rangers will do "a little extra damage" so when they go to play they make a ranger that has ONE interrupt power, and ONE of the various ways to add bonus damage, and they don't stack it all up because they just don't think about it. They think "this is the ordinary way people will play, they will want other kinds of feats and power, they won't take ALL the damage bonuses and interrupts" and you end up with 4e's issues (or 3.x's different issues, etc).
Ideally playtesting might find these issues, but most of the time the playtesters are building one-shot characters using half-finished rules and its HARD for them to test the totality of the game and see how it will play through in a campaign with players that will try to bend the system or just do what seems logical and take all the good stuff.
This is really why the total ditching of 4e frustrates me, because every time you go back and redo the whole core of the game you will just create this same pattern of issues in design and you won't probably end up with something better, because you've lost the benefit of all the insight you gained in the old system. WotC should be kicked in the head for the way they built 5e, it will suffer just the same as 3.5 and 4e have. They need to incrementally improve one game, not write a whole new one every 5 years.
